Peptidase Modulation of Vasoactive Intestinal Peptide PulmonaryRelaxation in Tracheal Superfused Guinea Pig LungsCraig M. uily,* Milton A. Martins,* and Jeffrey M. Drazen***Combined Program in Pulmonary and Critical Care Medicine, Departments of Medicine, Beth Israel Hospital and Brigham andWomen's Hospital, and Harvard Medical School; Ina Sue Perlmutter Laboratory, Children's Hospital; andtRespiratory Biology Program, Harvard School of Public Health, Boston, Massachusetts 02215

Abstract Introduction

The effects of enzyme inhibitors on vasoactive intestinal pep-tide (VIP)-induced decreases in airway opening pressure (Pao)and VIP-like immunoreactivity (VIP-LI) recovery were studiedin isolated tracheal superfused guinea pig lungs. In the absenceof inhibitors, VIP 0.38 (95% CI 033-0.54) nmol/kg animal,resulted in a 50%decrease in Pao and 33%of a 1 nmol/kg VIPdose was recovered as intact VIP. In the presence of two combi-nations of enzyme inhibitors, SCH32615 (S, 10 ;M) and apro-tinin (A, 500 tyrpsin inhibitor units ITIUJ/kg) or Sand soybeantrypsin inhibitor (T, 500 TIU/kg), VIP caused a significantlygreater decrease in Pao and greater quantities of VIP wererecovered from lung effluent (both P < 0.001). The addition ofcaptopril, (3 pM), leupeptin (4 pM), or bestatin (1 ;sM) failedto further increase pulmonary relaxation or recovery of VIP-LI. Whengiven singly, A, T, and S did not augment the effectsor recovery of VIP. The efficacy of S (a specific inhibitor ofneutral endopeptidase INEPI) and A and T (serine proteaseinhibitors) thus implicated NEPand at least one serine pro-tease as primary modulators of VIP activity in the guinea piglung. Wesought to corroborate this finding by characterizingthe predominant amino acid sites at which VIP is hydrolized inthe lung. When [mono('1I)iodo-Tyr'jVIP was offered to thelung, in the presence and absence of the active inhibitors, cleav-age products consistent with activity by NEPand a tryptic en-zyme were recovered. These data demonstrate that NEPand apeptidase with an inhibitor profile and cleavage pattern compat-ible with a tryptic enzyme inactivate VIP in a physiologicallycompetitive manner. (J. Clin. Invest. 1993. 91:235-243.) Keywords: vasoactive intestinal peptide * SCH32615 - aprotinin -

soybean trypsin inhibitor * captopril * bestatin * leupeptin * neu-tral endopeptidase (NEP) - tryptase

Address correspondence to Jeffrey M. Drazen, Combined Program inPulmonary and Critical Care Medicine, Brigham and Women's Hospi-tal, 75 Francis Street, Boston, MA02115.

Receivedfor publication 15 April 1992 and in revisedform 29 July1992.

1. Abbreviations used in this paper: A, aprotinin; B, bestatin; C, capto-pril; CI, confidence interval; IC30, inhibitor concentration resulting in50% inhibition; L, leupeptin; NEP, neutral endopeptidase; Pao, backpressure resulting from tracheal perfusion; RP-HPLC, reverse-phaseHPLC; S, enzyme inhibitor SCH32615; T, soybean trypsin inhibitor,TFA, trifluoroacetic acid; VIP, vasoactive intestinal peptide; VIP-LI,VIP-like immunoreactivity.

Vasoactive intestinal peptide (VIP)' is a 28-residue amidatedcarboxyl-terminal peptide with pulmonary relaxant activity(1-3). In the lung, VIP is found in neurons located at the me-dial-adventitial junction of pulmonary and bronchial vessels(4, 5). VIP is also present in cells subtending the smooth musclelayer and in glands of airways, including small bronchi (6).Although VIP has been shown to relax constricted airways andto inhibit glandular secretion (2, 7-10), its potency appears tobe limited by enzymatic degradation in the pulmonary mi-croenvironment. In isolated guinea pig (2) and human pulmo-nary contractile tissues, enzyme inhibitors have been shown toenhance VIP-induced relaxation (3, 11, 12), but the enzymesystems responsible for limiting VIP-induced pulmonary relax-ation have not been identified. Neutral endopeptidase (EC3.4.24.11 [NEP]) and mast cell proteases have been shown, inisolated systems, to cleave VIP with kinetics consistent withactivity under physiological conditions (3, 12-17). Since endog-enously released VIP may become available on the airway epi-thelium, we used the tracheal superfused lung (18, 19) to exam-ine the physiologically relevant pathways responsible for theenzymatic inactivation ofVIP. Our data suggest that the pulmo-nary relaxant activity of VIP is limited by enzymatic cleavageby NEPand a second protease with a cleavage pattern andinhibitor profile consistent with a tryptic enzyme.

Methods

Tracheal superfusion. Tracheal superfusion was performed as previ-ously described (19). 123 male Hartley-strain guinea pigs, 330-410 gbody wt, were anesthetized with 65 mg/kg sodium pentobarbital givenby intraperitoneal injection. Whenan appropriate plane of anesthesiawas achieved, a tracheostomy was created and a 2-cm length of polyeth-ylene tubing (1.67 mmi.d., 2.42 mmo.d.) was placed in the trachea.The abdominal cavity was opened and 500 U of heparin were injectedinto the vena cava. 3 min later the abdominal aorta was severed and theguinea pig was exsanguinated. The thoracic cavity was widely openedand the heart and lungs were removed en bloc. The lungs were dis-sected free and placed in a 37°C, 100% relative humidity plexiglass box.The lungs were superfused with a phosphate-buffered physiological so-lution with the following composition (mM): 137 NaCa, 1.8 CaCa2,1.05 MgCl2, 2.68 Ka, 0.6 NaHCO3, 0.13 NaH2PO4, 0.896 Na2HPO4,pH 7.4. The perfusion buffer was warmed to 45°C and pumped at 5ml/min through a bubble trap before being cooled to 37°C and enteredinto the lungs through the tracheal cannula. Perfusate exited the lungsvia numerous small holes placed in the pleura and was collected on icein polypropylene tubes containing glacial acetic acid (5% by final vol-ume) and was analyzed as indicated below. The "back pressure" result-ing from tracheal perfusion (Pao) was recorded from a side tap at thetracheal cannula using a pressure transducer (P23Db; Statham Instru-ments Inc., Oxnard, CA). Wehave previously shown that, at continu-ous flow, Pao reflects the contractile state of the lung (19). All experi-ments involving animals or tissues were approved by the animal careand use committees of both the Beth Israel and Children's Hospitals.

Determination of the time course and dose response of VIP-induced

Peptidase Modulation of Vasoactive Intestinal Peptide Pulmonary Relaxation 235

J. Clin. Invest.© The American Society for Clinical Investigation, Inc.0021-9738/93/01/0235/09 $2.00Volume 91, January 1993, 235-243

pulmonary relaxation. At the start of superfusion Pao rose slowly andreached a stable plateau after 15 min. Preliminary studies indicatedthat when airway tone was increased, by adding I0-' Mmethacholineto the perfusion buffer, VIP caused rapid decreases in Pao, hence alldata are reported in lungs with methacholine-induced airway tone. Toestablish the time course of VIP-induced relaxation, two groups of fivelungs were tracheally injected with either 1 nmol/kg VIP (all doses weregiven per kilogram animal) (diluted in 100 AI perfusion buffer) or 100Al of buffer alone, while Pao was recorded over a 15-min observationperiod. To establish a dose-response relationship for Pao as a functionof VIP dose, five lungs were tracheally injected with geometrically in-creasing doses of VIP, from 10 pmol to 100 nmol, given at 5-minintervals. For each lung, maximal relaxation was determined by tra-cheally injecting 10 nmol/kg of isoproterenol; a dose of isoproterenolshown in preliminary studies to elicit a maximal fall in Pao.

Perfusion with protease inhibitors. Ten groups of five animals eachwere superfused as described above; when a stable Pao had beenachieved I nmol/kg VIP or an equivalent volume of buffer was tra-cheally injected. VIP was administered with or without enzyme inhibi-tors, alone or in combination, added to the tracheal perfusate in thefollowing concentrations: captopril (C), 3 AM; aprotinin (A), 500 tryp-sin inhibitor units (TIU) (1 TIU will decrease the activity of two trypsinunits by 50% where one trypsin unit will hydrolyze 1 gmol of Na-ben-zoyl-DL-arginine p-nitroanilide/min at pH = 7.8 and 25°C) per kilo-gram animal; bestatin (B), 1 AM; leupeptin (L), 4 ,M; SCH32615 (S),10 uM; and soybean trypsin inhibitor (T), 500 TIU/kg animal. Groupsof five animals each were superfused under the following conditions:no inhibitors and no VIP injection (NO VIP); no inhibitors and a VIPinjection (VIP); single inhibitors and a VIP injection: S, A, T; andcombinations of inhibitors and a VIP injection: CBLS, CABLS,CTBLS, AS, and TS.

Recovery of VIP-like immunoreactivity. Lung effluent, from thesame lungs described above in the "Perfusion with enzyme inhibitors"section was assayed for VIP-like immunoreactivity. After administra-tion of 1 nmol/kg VIP, a 5-min fraction of lung effluent was collectedon ice in a polypropylene tube containing 5% by final volume glacialacetic acid. This fluid was processed as described below.

VIP immunoassay sample preparation. Lung effluent containing5%acetic acid was applied to a Sep-Pak® C,8 cartridge (Millipore Corp.,Bedford, MA) which was prewashed with 20 ml methanol followed by20 ml distilled H20. In the time course experiment the cartridges wereeluted with 3 ml of 60%acetonitrile and 40% H20 with 0. 1%trifluoro-acetic acid (TFA) titrated to pH 2.5 with triethylamine; in the inhibitorstudy the cartridges were eluted with 100% methanol. The cartridgeeluate was dried under vacuum (Savant Instruments Inc., Farming-dale, NY) and stored at -20°C until resuspended in 100 MI of a 50%methanol, 50% 0.1 N HCI solution for immunoassay.

Immunoassay of VIP. VIP was measured in lung effluent by solid-phase ELISA, modified from the method described by Folkesson et al.(20) for substance P. VIP ELISA plates were made by incubating VIP (5ng per well) conjugated to bovine serum albumin (The VIP-BSA conju-gate was made by adding 2.1 mgguinea pig VIP, 20,000 dpm '251-VIP,and 10.5 mgBSAto 735 Ml of PBSsolution, pH = 7.4, on ice. 360 Ml of25% glutaraldehyde in HPLCgrade H20 was added slowly and theresulting solution was incubated for 30 min at 0°C for 2 h at roomtemperature and was then extensively dialyzed against 0.01 MPBS.) inflat-bottomed polyvinyl microtiter plates. After incubation for 4 h atroom temperature, the plate was washed once in 0.15 MNaCl contain-ing 0.05% Tween 20, blocked with 3%normal goat serum (200 MI perwell), and incubated at room temperature for 1 h. The plate was washedthree times with 0.15 MNaCl containing 0.05% Tween 20 and oncewith distilled water. After drying the plates were wrapped in Parafilm0(American National Can Co., Greenwich, CT) and stored at -20°Cuntil use. At the time of assay, samples and VIP standards (3 x IO-7-10- " M) were dissolved in a solution of 50%methanol, 50%0.1 NHCI;25 Ml of each solution was added to a tube containing 200 Ml of theprimary antibody solution. The primary antibody solution contained

175 Ml anti-VIP antiserum (RAS 7161 N; Peninsula Laboratories Inc.,Belmont, CA), 600 MAl normal goat serum, 4 ml 0.5 MPBScontaining0.25% Tween 20, pH = 7.4, and 15.4 ml distilled H20. After I h, 100,Mlof each solution was transferred to a VIP ELISA plate well and incu-bated at 4°C for 72 h. The plates were then washed three times in 0.15MNaCl containing 0.05% Tween 20 and incubated for 2 h at 37°Cwith horseradish peroxidase-linked anti-rabbit antibody (NA934;Amersham International, Amersham, UK). The plates were washedthree times in 0.15 MNaCl with 0.05% Tween 20, developed withp-phenylenediamine substrate for 30 min, and the reaction wasstopped by adding 100 ,l 2.5 MH2SO4. The absorbance of each wellwas measured at 492 nm(Dynatech Laboratories, Inc., Chantilly, VA).The assay's lower limit of detection was 3 fmol of VIP; we found< 0.1% cross-reactivity to VIP 1-22, VIP 5-22, VIP 1-14, VIP 5-14, andVIP 13-28. VIP 5-28 had < 5% cross-reactivity with VIP 1-28.

Time courses of VIP recovery. To determine the time course ofrecovery of injected VIP and VIP cleavage products in lung effluent,three lungs were superfused and tracheally injected with 125,000 dpmof '251-VIP while 1-min fractions were collected; recovery of radiolabelwas assessed by gamma counting these fractions. To determine therecovery of intact VIP, three lungs were superfused and tracheally in-jected with 1 nmol/kg VIP while 1-min fractions of lung effluent werecollected for immunoassay. To determine the physiological activity ofVIP fragments, groups of three animals were superfused and tracheallyinjected with geometrically increasing doses, from 100 pmol/kg to 1,mol/kg, of VIP fragments and Pao was measured.

Recovery of 25I- VIP hydrolysis fragments from lung effluent. Ra-diolabeled VIP was perfused through the lung, effluent was collected,and radiolabeled fragments were resolved by reverse-phase HPLC(RP-HPLC). Retention times were compared with those of radiolabeledpeptide standards with sequences that would result from hydrolysis ofVIP at sites preferred by recombinant or highly purified NEP, chymase,and tryptase. Lungs were superfused as described above and after Paohad stabilized, a 5-min pre-injection control sample of lung effluentwas collected on ice in 5%glacial acetic acid. A second 5-min fractionof lung effluent was collected after tracheal injection of 250,000 dpmof[mono('251I)iodo-Tyr'0]VIP. Samples were dried under vacuum, resus-pended in 600 Ml of HPLCgrade H20 with 0.1 %(vol/vol) TFA, dividedinto two 300-Ml aliquots, passed through a 0.22-,m filter (MilliporeCorp., Bedford, MA), and subjected to RP-HPLC analysis. Lung ef-fluent was analyzed from four groups of two lungs superfused withbuffer containing the following inhibitor combinations (at concentra-tions defined above) added to the superfusate: CBL, CABL, CBLS, andCABLS.

Preparation of [mono('251)iodo-Tyr']VIP and [('251)iodo-Tyr'0]-VIPfragments. VIP or VIP fragments were iodinated by the method ofMartin et al. (21). 1.5 nmol of VIP or a VIP fragment was dissolved in25,Ml of 0.3 MPBS, pH = 7-11, and 1 mCi of carrier-free sodium iodide(0.5 nmol in 10 Ml NaOH, pH = 7.4; Amersham International), wasadded, providing 3 molecules of VIP or VIP fragment for each 125Iatom. While stirring, iodination was initiated with the addition of 14.2nmol of chloramine-T at a concentration of 1 mg/ml. After 15 s atroom temperature the reaction was terminated by adding 42.1 nmol ofNa2S2SO5. The reaction mixture was applied to a Sep-PakO Cl8 car-tridge and washed with 9 ml of HPLC grade H20 with 0.1% TFA.1251-VIP or '251-VIP fragments were eluted in a solution of 60% acetoni-trile and 40% H20 with 0.1 %TFA (pH 2.5). The eluate was purified byC,8 reverse-phase chromatography using the RP-HPLC protocol de-scribed below as protocol I. Two sequential HPLC purification stepsyielded a material with a single peak and a reproducible retention time.Purified material was stored at 4°C and was used the next day.

Identification of VIP and VIP hydrolysis fragments in lung effluent.Identification of potential VIP cleavage products that would be formedby action of NEP, chymase, or tryptase, (13, 14), i.e., VIP 1-22, VIP1-2 1, VIP 1-14, VIP 4-14, VIP 5-14, VIP 5-21, VIP 5-22, VIP 5-28, VIP9-10, VIP 10-1 1, VIP 21-22, VIP 22-23, and VIP 22-28, was accom-plished by comparing the RP-HPLC retention time of authentic 125I-

236 C. M. Lilly, M. A. Martins, and J. M. Drazen

VIP and iodinated synthetic fragments with the retention time of '25Iradiolabel recovered from lung effluent. Coelution of radiolabel withsynthetic standard on two distinct HPLCprotocols was required forradiopeptide identification. RP-HPLC separation was accomplishedusing a programmable multiwavelength detector (model 490) pump(model 510) and a 3.9 x 300 mm, C18 reverse-phase column (Nova-PakO) (all from Waters Instruments, Milford, MA) at a flow rate of 1ml/min. The initial mobile phase for both HPLCprotocols was HPLCgrade H20 with 0.1% TFA. In elution protocol 1, the column waseluted with a 0.7-49% linear gradient of acetonitrile over 40 min. Inelution protocol 2, the column was elpted with a 14-49% linear gra-dient of acetonitrile over 55 min. 30-s fractions were collected and 1251content determined by gamma counting (Packard Instrument Co.,Inc., Sterling, VA). Authentic standards for all peptides were processedthrough the system without a lung present; the recovery among thevarious peptide was 33-95%; radiochromatograms were corrected forthe background activity and for losses due to processing. The inhibitorcombinations used had no effect on the elution time of the VIP break-down products.

Materials. VIP (guinea pig) 1-28, VIP fragments 1-21, 5-14, 5-21,5-22, 5-28, 9-10, 10-11, 21-22, 22-23, 22-28, and VIP (guinea pig)antisera were obtained from Peninsula Laboratories Inc.). VIP frag-ments 1-22, 1-14, and 5-22 were custom synthesized by Hospital Pep-tide Services (Boston, MA). VIP 4-14 was custom synthesized by Re-search Genetics (Huntsville, AL). Glutaraldehyde, captopril, bestatin,leupeptin, aprotinin, soybean trypsin inhibitor, chloramine-T, metabi-sulfite, and methacholine were obtained from Sigma Chemical Co. (St.Louis, MO). SCH 32615 was the kind gift of Schering Corporation(Kenilworth, NJ). All other materials were HPLCor reagent grade asappropriate.

Statistical analysis. All values are expressed as mean±95% confi-dence intervals (CI), unless otherwise stated. The statistical significanceof differences between means was determined by analysis of variance. P< 0.05 was considered significant. Whenanalysis of variance indicateda significant difference the Newman-Keuls test was used to determinewhich groups were significantly different from each other.

Results

Pulmonary relaxationDose-response relationship of VIP on pulmonary relaxation.Fig. 1 presents the dose-response relationship for VIP on Pao,measured in the tracheal superfused guinea pig lung. Increasingdoses of VIP were associated with a progressive fall in Pao. TheID50 was 0.38 nmol/kg (95% CI, 0.33-0.54). The superfusedlung VIP ID50 of 0.38 nmol/kg equates to an approximate IC50of 12.6 nM, assuming an approximate dilution volume of 30ml, which is similar to the reported tissue bath IC50 for VIP (2,10,22,23). The maximal relaxation after 100 nmol/kg VIP was52 cm H20 (95% CI, 34-71). The maximal relaxation after 10nmol/kg isoproterenol was 56 cm H20 (95% CI, 37-74); com-pared with the fall in Pao after 100 nmol/kg VIP this differencewas not significant (P = 0.79). None of the VIP fragments hadpulmonary relaxant activity. Fig. 3 A presents the effect of VIPon the maximal fall in Pao. Comparison of results from thegroup receiving only VIP (VIP) to the group not receiving VIP(NO VIP) demonstrates that VIP injection was associated witha significantly greater maximal decrease in Pao, 33 cm H20(95% CI, 24-43) in the VIP group and 3 cm H20 (95% CI,-6-13) in the NOVIP group. A significantly greater fall in Paowas observed in all groups receiving VIP compared with theNOVIP group (P < 0.001).

Time course of VIP-induced changes in Pao. The timecourse and magnitude of changes in Pao after tracheal injection

0Io Za-.

-(-)

w) 0

< C)oC)cLL 0

c,, d)

0 lOpmol lOOpmol 1nmol lOnmol lOOnmol MaximalIsoproterenol

VIP DOSE/Kg Animal Response

Figure 1. Airway opening pressure as a function of geometrically in-creasing doses of tracheally injected VIP. The ID50 was 0.38 (95% CI,0.33-0.54) nmol/kg VIP. Data are presented as the group mean with95% CI, n = 5.

of saline or 1 nmol/kg VIP and the time course of appearanceof VIP-like immunoreactivity (VIP-LI) in lung effluent areshown in Fig. 2. Tracheal infusion of VIP was associated with asignificant (P < 0.0001) decline in Pao. The time for 50% relax-ation was 1.5 min (95% CI, 0.9-2.1). Maximal relaxation wasachieved by - 4 min and Pao returned toward baseline over 15min. Similar time courses were observed in lungs treated withenzyme inhibitors (data not shown).

Effects of active enzyme inhibitors on Pao. Wefound thatthe TS and AS combinations significantly enhanced VIP-in-duced pulmonary relaxation. In the presence of TS, VIP injec-tion resulted in a Pao decrease of 55 cm H20 (95% CI, 45-64),whereas the presence of AS was associated with a Pao decreaseof 61 cm H2O(95% CI, 52-71). The changes in Pao for the ASand TS groups were significantly different from those in theVIP, S, T, A and CBLS groups, P < 0.001, for each com-parison.

Effects of single inhibitors on Pao. There was no significantdifference in the Pao response between the VIP group (withoutinhibitors) and the groups receiving only one of the inhibitors.Pao decreased 27 cm H20 (95% CI, 18-37) in the A group, 34cm H20 (95% CI, 44-25) in the S group, and 25 cm H20 (95%CI, 15-34) in the T group, all P = NS compared with the noinhibitor (VIP) group.

Effects of combined inhibitors on Pao. In the absence of VIPinfusion, the inhibitors did not induce a decrease in Pao (datanot shown). WhenCBLwas added to AS or TS, giving CABLSor CTBLS, no further enhancement of VIP-induced pulmo-nary relaxation was observed. Pao decreased 53 cm H20 (95%CI, 62-44, P < 0.001) in the CABLSgroup and decreased 59cm H20 (95% CI, 50-69, P < 0.001) in the CTBLS group.Although not significantly different from the AS or TS groups,the changes in Pao for the CABLSor CTBLSgroups were signif-icantly different than the VIP, S, T, A, and CBLS groups, P< 0.01 for each comparison. The changes in Pao in the CBLSgroup, which was treated with a combination of inhibitors thatcontained all of the inhibitors except aprotinin and soybeantrypsin inhibitor, was not significantly different from the VIPonly group. In this CBLS group, Pao decreased 33 cm H20(95% CI, 23-42) P = NS compared with the VIP and singleinhibitor groups.

Peptidase Modulation of Vasoactive Intestinal Peptide Pulmonary Relaxation 237

p<.0001I

-3 -2 -1 0

Trac heaInjection

(lnmol/kg VIP)

I I I I9 I1 2 3 4 5 6 7 8 9 10

TIME (Minutes)

>

, C

Za

m 1'

_tnw

Y aK

>-

l10-

10 I-

p'O.O1

.

lT

7 8011 2 3 4 5 6

TrachealInjection

(1 nmol/kg VIP)

TIME (Minutes)

Recovery of VIP-like immunoreactivity

Time course of VIP-induced changes in VIP recovery. The timecourse and magnitude of changes in the appearance of VIP-LIin lung effluent are shown in Fig. 2. Tracheal infusion of 1

nmol/kg VIP was associated with a significant (P < 0.01) in-crease in VIP-LI detected by immunoassay in 1-min fractionsof lung effluent. When 5-min fractions of lung effluent were

immunoassayed, VIP-LI increased from 1.6 (95% CI, 0-5.3) to30 pmol (9.3-51) after tracheal injection of 1 nmol/kg VIP (P< 0.01); 33±4.6% (mean±SEM) of the injected VIP dose was

recovered as VIP-LI. This is similar to the data presented inFig. 3, which shows a significant increase in VIP-LI in allgroups receiving VIP compared with the NOVIP group (P< 0.001). The total recovery of radiolabel after tracheal injec-tion of 1251I-VIP was 94±3%; this was not significantly differentfrom the recovery of 1251 when '251-VIP was injected into theperfusion apparatus without a lung present.

Effects of active inhibitors on VIP-LI recovery. Wefoundthat the TS and AS combinations significantly increased VIP-LI recovery from lung effluent. The changes for the ASand TS

Figure 2. (A) Time course of the decrease in Paoafter tracheal injection of I nmol/kg VIP. Themean time for 50% relaxation was 1.5 (95% CI,0.9-2.1) min. Data are presented as the group

mean with 95% CI, n = 5, *P < 0.0001 for lungsreceiving VIP (solid circles) compared withbuffer-injected control lungs not receiving VIP(open circles). (B) Time course of recovery ofVIP-LI from lung effluent after VIP tracheal in-jection. Data are presented as the mean andSEM, n = 3, TP < 0.01 compared with lung ef-fluent VIP-LI before VIP injection.

groups were significantly different from those in the VIP, S, T,A, and CBLSgroups, P < 0.001 for each comparison.

Effects of single inhibitors on VIP-LI recovery. There was

no significant difference in lung effluent VIP-LI between theVIP group (without inhibitors) and the groups receiving onlyone of the inhibitors. Lung effluent from the A group con-

tained 13 pmol (95% CI, 5-21), the S group contained 14 pmol(95% CI, 6-22), the T group contained 13 pmol (95% CI, 5-21),and the no inhibitor VIP group contained 18 pmol (95% CI,10-26), all P = NS.

Effects of combined inhibitors on VIP-LI recovery. WhenCBL was added to AS or TS, giving CABLS or CTBLS, no

further enhancement of peptide recovery was observed. Al-though not significantly different from the AS or TS groups,the changes in the CABLS and CTBLS groups were signifi-cantly different from the VIP, S, T, A, and CBLS groups, P< 0.01 for each comparison. The changes in Pao in the CBLSgroup, which was treated with a combination of inhibitors thatcontained all of the inhibitors except aprotinin and soybeantrypsin inhibitor, was not significantly different from the VIP

only group. In this CBLSgroup 26 pmol VIP-LI (95% CI, 18-

238 C. M. Lilly, M. A. Martins, and J. M. Drazen

A

0-

0cwo

I t( a)

° c

z- a.

L o

Z

%-

I (

0-

-10 -

-20-

-30-

-40

B

No NoVIP Inhibitor CBLS S

w

0 o

c) 0

- .2_cL

w

IC>j

No No CBLS S A T TS CTBLS AS CABLSVIP Inhibitor

34) was recovered from lung effluent, both P = NScomparedwith the VIP and single inhibitor groups.

Identification of 2I- VIP hydrolysisfragments. Our data on

the magnitude of physiological response and recovery of VIP-LI were consistent with enzymatic degradation as limiting thepulmonary relaxant activity of VIP. Furthermore, the require-ment for two enzyme inhibitors suggested that NEP, as inhib-ited by its specific inhibitor SCH32615, and enzymes inhibitedboth by aprotinin and soybean trypsin inhibitor were the en-

zyme systems limiting VIP activity. To confirm that VIP pul-monary relaxant activity is limited by enzymatic degradationand to identify the physiologically relevant enzyme systems, we

identified the major breakdown products of 1251I-VIP in super-

fused lung effluent. Further, we examined the effects of thephysiologically active enzyme inhibitors on the qualitative re-

covery of these '251I-VIP breakdown products. Resolution of therecovered radiolabel by RP-HPLC allowed separation of VIPfrom its enzymatic cleavage products and fragment identifica-tion was accomplished by co-chromatography with syntheticstandards (Fig. 4). When both physiologically active enzyme

inhibitors were present, the percentage of total counts recov-

ered that coeluted with intact VIP increased from 29 to 68%.The magnitude of this increase is similar to the increase in

Figure 3. (A) Pao response after tracheal injectionof 1 nmol/kg VIP. Data shown are means with95%CI for groups of five lungs treated with singleinhibitors designated by single letters, with combi-nations of inhibitors designated by combinationsof letters, or control lungs.

VIP-LI observed with these same inhibitors present (Fig. 3).Once we confirmed that the physiologically active inhibitorsincreased the recovery of intact VIP from lung effluent anddecreased the fraction of the radiolabel coeluting with VIPcleavage products, we identified and quantified the specificcleavage products present in lung effluent. Wewere able toimplicate specific enzyme systems by comparing the observedVIP breakdown products with VIP cleavage products deducedfrom known VIP cleavage sites (Fig. 5). Whenneither S nor Awas present, 29% of the radiolabel coeluted with intact VIP.40% of the radiolabel coeluted with VIP 4/5-14 (VIP 4-14 or

VIP 5-14), 16% of the radiolabel coeluted with VIP 1-14, and7%of the counts coeluted with VIP 5-21/22 (VIP 5-21 or VIP5-22) (Fig. 6). Wewere able to account for 92%of the adminis-tered radiolabel. When S was present and A was not present,26%of the radiolabel coeluted with intact VIP. The percentageof the radiolabel coeluting with VIP 4/5-14, products of NEPand tryptic cleavage, decreased to 24%. The percentage of theradiolabel coeluting with the tryptic product VIP 1-14 in-creased to 43%. Wewere able to account for 93%of the admin-istered radiolabel. WhenA was present and S was not present,24% of the radiolabel coeluted with intact VIP. 70% of theradiolabel coeluted with VIP 5-21/22. Wewere able to account

Peptidase Modulation of Vasoactive Intestinal Peptide Pulmonary Relaxation 239

ALU

wcn

cra-0z

W .

0 !

> 3:

z

w0zIC-)

B

0 10 20 30 40

Time (Minutes)

VIP 1-28

HPLC Protocol 1

40

Figure 4. RP-HPLC resolution of intact '25I-VIP1-28 from '251-VIP cleavage products in lung

50 effluent, identified as '25I-VIP 4/5-14 and 12511VIP 1-14. (A) In the absence of and (B) in thepresence of SCH32615 and aprotinin.

Mast Cell Tryptase Mast Cell Tryptase

1251r2sI ~~~~~~~~IFIF-H

HIS -SER-ASP-ALA-LEU-PHE-THR-ASP-THR-TYR-THR-ARG-LEU-ARG -LYS-GLN-MET-ALA-MET- LYS- LYS-TYR-LEU-ASN-SER-VAL-LEU-ASN -NH2

1 2 3j4j5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21j22 23 24 25 26 27 28

Neutral Endopeptidase Neutral Endopeptidase(EC 3.4.24. 11)

Mast Cell Chymase

Figure 5. The most active cleavage sites for VIP (guinea pig) for the limited number of enzymes known to cleave VIP under physiological con-ditions are shown.

240 C. M. Lilly, M. A. Martins, and J. M. Drazen

A

'/,

c00-C

E0.(

B

50

60 F

50 F

40

30h

c0O=30-

HEa.u 20h

10Fo

VIP 1-14

VIP 4,5-14

....1. 4N

0 10 20 30

Time (Minutes)

CONDITION 1>, HPLC PROTOCOL1 40% >. HPLC PROTOCOL2WL 4-14 29% J 32%2%

40 22% C 2P20 T- 1-28 0 7% 16% 1H28

J 5i21 I 20 - 5-21 Mcr -2~~I II 8%5-22 14I

25min 25minHPLC ELUTION TIME HPLC ELUTION TIME

(1.5 min intervals) (1.5 min intervals)

CONDITION 2> PROTOCOL > PROTOCOL2 Ccr40 ~~~36% cr40 -

~~~~~25% 5-21 26% >

Lu20-1-4 ~ 2 45-2JFl1-2828

25min 25minHPLC ELUTION TIME HPLC ELUTION TIME

(1.5 min intervals) (1.5 min intervals)

CONDITION 3

>-80 PROTOCOL 76% >- 80-PROTOCOL2 70%x cr~~~~~~~~~~~i-52> 5-22 > ~~~~~~~~~~~~~~~~~~~~~~5-22LAJ LL) ~ ~ ~ ~ ~ ~~~1-424%

25min 25minHPLC ELUTION TIME HPLC ELUTION TIME

(1.5 min intervals) (1.5 min intervals)

for 94% of the administered radiolabel. VIP 1-22 (a potentialchymase product) was not identified under any of the condi-tions studied.

Discussion

Our data demonstrate that combinations of the specific NEPinhibitor SCH32615 (24) and a serine protease inhibitor, apro-tinin or soybean trypsin inhibitor, significantly increase VIP-induced pulmonary relaxation and the recovery of intact VIP.The same enzyme inhibitor combinations enhance recovery ofintact VIP from lung effluent. Taken together these findingssupport the hypothesis that the pulmonary relaxant activity ofVIP is limited by enzymatic degradation by more than oneenzyme system. The relevant enzyme systems are likely to beNEP and another enzyme system whose actions are limitedboth by aprotinin and soybean trypsin inhibitor.

Aprotinin and soybean trypsin inhibitor are broad spec-trum inhibitors (2, 3), each can inhibit serine proteases, includ-ing a number of tryptic enzymes potentially found in the lung.Tryptase and chymase are the only enzymes for which VIPcleavage sites have been reported (Fig. 5) that may also be sensi-tive to both aprotinin and soybean trypsin inhibitor (25, 26).We isolated radiolabeled VIP cleavage fragments from lungeffluent and identified them as VIP 5-21/22 and VIP 1-14;these are expected VIP cleavage products for NEPand a trypticenzyme. The observed changes in the VIP fragment profile, inthe presence and absence of enzyme inhibitors, can be fullyexplained by inhibition of NEPand a tryptic moiety; we did

Figure 6. '231-VIP and 12511VIP breakdown productsidentified by cochromato-graphy with synthetic stan-dards on two HPLCproto-cols. In condition 1, withneither S nor A present,28% of the radiolabel co-eluted with '251-VIP, 40%coeluted with 125I-VIP 4/5-14 (a product of NEPandtryptase cleavage), and 16%coeluted with 125I-VIP 1-14(a tryptase product). Incondition 2, with S presentand A not present, 25%ofthe radiolabel coeluted with125I-VIP, while the activitycoeluting with 1251-VIP 1-14(the tryptase product) in-creased to 43% and activitycoeluting with '251-VIP 4/5-14 decreased to 24%. Incondition 3, with S notpresent and A present, 24%of the radiolabel coelutedwith 1251-VIP and 70%ofthe radiolabel coeluted with'251-VIP 5-21/22, a productof NEPcleavage.

not identify any VIP cleavage products that could only be ex-plained by chymase activity. Whenboth SCH32615 and apro-tinin were present in the perfusate (Fig. 4), the recovery ofintact VIP (1-28) increased and the recovery of VIP cleavageproducts (VIP 4/5-14 and 1-14) decreased compared with VIPrecovery in the absence of these enzyme inhibitors. Further-more, the magnitude of increase of intact '25I-VIP recovery wassimilar to the increase in VIP-LI when inhibitor combinationAS was present, confirming the results of the ELISA study.Whenneither aprotinin nor SCH32615 was present (condition1, Fig. 6), products consistent with tryptic activity (VIP 1-14)and combined NEP and tryptic activity (VIP 4/5-14) wereidentified in lung effluent. When NEPwas inhibited by SCH32615 but aprotinin was absent, (condition 2, Fig. 6), there wasdecreased recovery of products consistent with NEP activity(VIP 4/5-14) and increased recovery of VIP 1-14, a productconsistent with tryptic activity. When SCH32615 was absentfrom and aprotinin was present in the perfusion buffer (condi-tion 3, Fig. 6), the recovery of products consistent with NEPactivity increased and the product consistent with tryptic activ-ity was absent. In each condition, the changes in VIP productrecovery profile are those that would be expected from inhibi-tion of NEP (SCH 32615) and a tryptic enzyme (aprotinin)whose identity remains unknown but whose cleavage site is atthe VIP 14-15 bond. In each case the radiolabeled fragmentsaccounted for > 90% of the radiolabeled peptide offered to thelung. In addition to implicating NEP and a tryptic enzyme,these data diminish the potential physiological importance ofangiotensin-converting enzyme, aminopeptidases, and car-

Peptidase Modulation of Vasoactive Intestinal Peptide Pulmonary Relaxation 241

boxypeptidases in limiting VIP activity. The failure of L toenhance VIP recovery or decrease in Pao suggests that it is notan effective inhibitor of guinea pig tryptic enzymes. Since thecleavage products were shown to be physiologically inactiveand the presence of intact '25I-VIP or VIP-LI in lung effluentcorrelates with VIP-induced pulmonary relaxation, these find-ings strongly support the hypothesis that the magnitude of VIP-induced pulmonary relaxation is limited by enzymatic degrada-tion.

Our data are consistent with NEPand a tryptic enzyme asthe physiologically relevant systems responsible for VIP inacti-vation. NEPis known to have an important regulatory role inthe pulmonary microenvironment. In the lung, NEPhas beenlocalized to epithelial cells and airway smooth muscle cells(27-29). It is well established that the bronchoconstrictor neu-ropeptides substance-P and neurokinin-A have their physiologi-cal activity limited by enzymatic degradation at or near theirsite of release or action. Neurokinin-A pulmonary activity islimited by NEP (19, 30), whereas SP is limited by both NEPand angiotensin-converting enzyme. VIP is a known substratefor NEP (Fig. 5) (13, 31). However, despite the availability ofspecific inhibitors it has been difficult to demonstrate a role forNEP in regulating VIP physiological activity. When VIP-in-duced relaxation in guinea pig (2) and human (3) tracheal ringswas studied in the presence of enzyme inhibitors, inhibitor com-binations that combined a NEP inhibitor with aprotinin or

soybean trypsin inhibitor augmented the relaxant effects ofVIP; single inhibitors had little physiological effect. Therefore,even though NEP activity has the capacity to limit VIP-in-duced pulmonary relaxation, other enzyme systems are physio-logically competitive with NEP.

Among tryptic enzymes known to be resident in lung tis-sues, tryptase is stored in and released from mast cells (32) andhas been implicated (33), but not widely recognized, as a modu-lator of peptide activity. Since a proportion of mast cells islocated at the site of peptidergic nerve terminals (34-36), suchcells are uniquely poised to exert a regulatory role in this mi-croenvironment. VIP is known to be a substrate for tryptase(14) and the addition of tryptase has been shown to reverseVIP-induced tracheal relaxation in vitro (37). Weidentified, inlung effluent, VIP cleavage products that would be expected toresult from tryptase activity and combined tryptase and NEPactivity. Taken together these findings are consistent with tryp-tase as the tryptic enzyme that modulates VIP-induced pulmo-nary relaxation. This is a striking finding since by enzyme histo-chemistry guinea pig mast cells contain virtually no trypticactivity (33, 38). It is possible however that tryptase is notstored in guinea pig mast cells in the concentrated form de-tected by this method. It is also well established that mast celltryptase activity likely represents action of a family of mast cellproteases with related but distinct amino acid sequences, sub-strate specificity, inhibitor sensitivity, and immunoreactivity(39, 40). Our data support the presence of a physiologicallyactive enzyme consistent with mast cell tryptase in the guineapig lung.

In summary, we have demonstrated that VIP-induced pul-monary relaxation is limited by enzymatic degradation by twoequally important enzyme systems that are physiologicallycompetitive with the VIP receptor for terminating the effects ofVIP. Our data support the identification of these two enzymesystems as NEPand a tryptic enzyme. These findings demon-

strate that a regulatory scheme, involving the simultaneous ac-tion of multiple enzymes limiting VIP pulmonary relaxation,which has been proposed by others (2, 3), is of physiologicalsignificance.

Acknowledgments

The authors thank Dr. Bernard Ransil for his helpful advice concern-ing statistical analysis. Data analysis was performed on the Core Labcomputer facilities of the Beth Israel Hospital.

This work was supported by National Institutes of Health grantHL-39827 and a grant from the Brazilian National Council for Scien-tific Development (CNPQ).

References

1. Said, S. I., and V. Mutt. 1969. Long acting vasodilator peptide from lungtissue. Nature (Lond.). 224:699-700.

2. Thompson, D. C., L. Diamond, and R. J. Altiere. 1990. Enzymatic modula-tion of vasoactive intestinal peptide and nonadrenergic noncholinergic inhibitoryresponses in guinea pig tracheae. Am. Rev. Respir. Dis. 142:1119-1123.

3. Tam, E. K., G. M. Franconi, J. A. Nadel, and G. H. Caughey. 1990.Protease inhibitors potentiate smooth muscle relaxation induced by vasoactiveintestinal peptide in isolated human bronchi. Am. J. Respir. Cell. Mol. Bio.2:449-452.

4. Dey, R. D., W. A. Shannon, and S. I. Said. 1981. Localization of VIP-im-munoreactive nerves in airways and pulmonary vessels of dogs, cats, and humansubjects. Cell Tissue Res. 220:231-238.

5. Dey, R. D., J. B. Altemus, and M. Michalkiewicz. 1991. Distribution ofvasoactive intestinal peptide- and substance-P-containing nerves originatingfrom neurons of airway ganglia in cat bronchi. J. Comp. Neurol. 304:330-340.

6. Laitinen, A., M. Partanen, A. Hervonen, M. Pelto-Huikko, and L. A. Lai-tinen. 1985. VIP like immunoreactive nerves in human respiratory tract; lightand electron microscopic study. Histochemistry. 82:313-319.

7. Saga, T., and S. I. Said. 1984. Vasoactive intestinal peptide relaxes isolatedstrips of human bronchus, pulmonary artery and lung parenchyma. Trans. Assoc.Am. Physicians. 97:304-3 10.

8. Palmer, J. B. D., F. M. C. Cuss, and P. J. Barnes. 1987. VIP and PHMandtheir role in nonadrenergic inhibitory responses in isolated human airways. J.Appl. Physiol. 61:1322-1328.

9. Coles, S. J., S. I. Said, and L. M. Reid. 1981. Inhibition by VIP of glycon-conjugate and lysozyme secretion by human airways in vitro. Am. Rev. Respir.Dis. 124:531-536.

10. Webber, S. E., and J. G. Widdicombe. 1987. The effect of vasoactiveintestinal peptide on smooth muscle tone and mucus secretion from the ferrettrachea. Br. J. Pharmacol. 91:139-148.

1 1. Farmer, S. G., and J. Togo. 1990. Effects of epithelium removal on relax-ation of airway smooth muscle induced by vasoactive intestinal peptide andelectrical field stimulation. Br. J. Pharmacol. 100:73-78.

12. Hachisu, M., T. Hiranuma, S. Tani, and T. Izuka. 1991. Enzymatic degra-dation of helodermin and vasoactive intestinal polypeptide. J. Pharmacobio-dyn.14:126-131.

13. Goetzl, E. J., S. P. Sreedharan, C. W. Turck, R. Bridenbaugh, and B.Malfroy. 1989. Preferential cleavage of amino- and carboxyl-terminal oligopep-tides from vasoactive intestinal polypeptide by human recombinant enkephalin-ase (neutral endopeptidase, EC 3.4.24.11). Biochem. Biophys. Res. Commun.158:850-854.

14. Caughey, G. H., F. Leidig, N. F. Viro, and J. A. Nadel. 1988. Substance Pand vasoactive intestinal peptide degradation by mast cell tryptase and chymase.J. Pharmacol. Exp. Ther. 244:133-137.

15. Powers, J. C., T. Tanaka, J. W. Harper, Y. Minematsu, L. Barker, D.Lincoln, K. V. Crumley, J. E. Fraki, N. M. Schechter, G. G. Lazarus, et al. 1985.Mammalian chymotrypsin-like enzymes. Comparative reactivities of rat mastcell proteases, human and dog skin chymases, and human cathepsin G withpeptide 4-nitroanilide substrates and with peptide chloromethyl ketone and sul-fonyl flouride inhibitors. Biochemistry. 24:2048-2058.

16. Tanaka, T., B. J. McRae, K. Cho, R. Cook, J. E. Fraki, D. A. Johnson, andJ. C. Powers. 1983. Mammalian trypsin-like enzymes. Comparative analysis ofhuman skin tryptase, human lung tryptase, and bovine trypsin with peptide 4-ni-troanilide and thioester substrates. J. Biol. Chem. 258:13552-13557.

17. Bodansky, M., Y. S. Klausner, and S. I. Said. 1983. Biologic activities ofsynthetic peptides corresponding to fragments of and to the entire sequence ofvasoactive intestinal peptide. Proc. Natl. Acad. Sci. USA. 70:382-384.

18. De Nucci, G., and S. Moncada. 1987. Release of vasoactive substances

242 C. M. Lilly, M. A. Martins, and J. M. Drazen

from guinea pig isolated lungs perfused via the trachea. Am. Rev. Respir. Dis.135:S39-S41.

19. Martins, M. A., S. A. Shore, N. P. Gerard, C. Gerard, and J. M. Drazen.1990. Peptidase modulation of the pulmonary effects of tachykinins in trachealsuperfused guinea pig lungs. J. Clin. Invest. 85:170-176.

20. Folkesson, R., A. Neil, and L. Terenius. 1985. Enzyme-linked immuno-sorbent assay of substance-P and its metabolite SP 1-7. A comparison with RIA.J. Neurosci. Methods. 14:169-176.

21. Martin, J. L., K. Rose, G. J. Hughes, and P. J. Magistretti. 1986.(mono('25iodo-Tyr'0,MetO'7)-Vasoactive intestinal polypeptide. J. Biol. Chem.261:5320-5327.

22. Stretton, D. C., M. G. Belvisi, and P. J. Barnes. 1990. Sensory depletionpotentiates inhibitory non-adrenergic, non-cholinergic nerves in guinea pig air-ways. Eur J. Pharmacol. 184:333-337.

23. Shikada, K., A. Yamamoto, and S. Tanaka. 1991. Effects of phosphodies-terase inhibitors on vasoactive intestinal peptide-induced relaxation of isolatedguinea pig trachea. Eur J. Pharmacol. 195:389-394.

24. Yaksh, T. L., M. B. Sabbe, D. Lucas, E. Mjanger, and R. E. Chipkin. 1991.Effects of (N4L-l-carboxyl-2-phenyl)ethyl))-L-phenylalanyl-beta-alanine (SCH32615) a neutral endopeptidase (enkephalinase) inhibitor, on levels of enkepha-lin, encrypted enkephalins and substance P in cerebral spinal fluid and plasma ofprimates. J. Pharmacol. Exp. Ther. 256:1033-1041.

25. Caughey, G. A., N. F. Viro, J. Ramachandran, S. C. Lazarus, D. B. Bor-son, and J. A. Nadel. 1987. Dog mastocytoma tryptase: affinity purification,characterization, and amino-terminal sequence. Arch. Biochem. Biophys.258:555-563.

26. Caughey, G. A., N. F. Viro, S. C. Lazarus, and J. A. Nadel. 1988. Purifica-tion and characterization of dog mastocytoma chymase: identification of an octa-peptide conserved in chymotryptic leukocyte proteinases. Biochim. Biophys.Acta. 952:142-149.

27. Matsas, R., A. J. Kenny, and A. J. Turner. 1984. The metabolism ofneuropeptides. The hydrolysis of peptides, including enkephalins, tachykininsand their analogues, by endopeptidase-24. 11. Biochem. J. 223:433-440.

28. Edros, E. G., and R. A. Skidgel. 1989. Neutral endopeptidase 24.11 (en-kephalinase) and related regulators of peptide hormones. FASEB(Fed. Am. Soc.Exp. Biol.) J. 3:145-151.

29. Ryan, J. W. 1989. Peptidase enzymes of the pulmonary vascular surface.Am. J. Physiol. 257:L53-L60.

30. Dusser, D. J., E. Umeno, P. D. Graf, T. Djokic, D. B. Borson, and J. A.Nadel. 1988. Airway neutral endopeptidase-like enzyme modulates tachykinin-induced bronchoconstriction in vivo. J. Appl. Physiol. 65:2585-2591.

31. Barbato, G. F., F. Jordan, and B. R. Komisaruk. 1988. The in vitroproteolytic processing of vasoactive intestinal polypeptide by rat spinal cord ho-mogenate. Ann. NYAcad. Sci. 527:582-585.

32. Schwartz, L. B., R. A. Lewis, D. Seldin, and K. F. Austen. 1981. Acidhydrolases and tryptase from secretory granules of dispersed human mast cells. J.Immunol. 126:1290-1294.

33. Caughey, G. H. 1989. Role of mast cell tryptase in airway function. Am. J.Physiol. 257:L39-L46.

34. Newson, B., A. Dahlstrom, L. Enerback, and H. Ahlman. 1988. Sugges-tive evidence for a direct innervation of mucosal mast cells. An electron micro-scopic study. Neuroscience. 10:565-570.

35. Stead, R. H., M. Tomioka, G. Quinonez, G. T. Simon, S. Y. Felten, and J.Bienenstock. 1987. Intestinal mucosal mast cells in normal and nematode in-fected rat intestines are in intimate contact with peptidergic nerves. Proc. Natl.Acad. Sci. USA. 84:2975-2979.

36. Weiner-Menzel, L., B. Schultz, F. Vakilzadeh, and B. M. Czarnetzki.1981. Electron microscopical evidence for a direct contact between nerve fibersand mast cells. Acta Dermato-venereol. 61:465-469.

37. Franconi, G. M., P. D. Graf, S. C. Lazarus, J. A. Nadel, and G. C.Caughey. 1989. Mast cell tryptase and chymase reverse airway smooth musclerelaxation induced by vasoactive intestinal peptide in the ferret. J. Pharmacol.Exp. Ther. 248:947-951.

38. Chiu, H., and D. Lagunoff. 1972. Histochemical comparison of vertebratemast cells. Histochem. J. 4:135-144.

39. Reynolds, D. S., D. S. Gurley, K. F. Austen, and W. E. Serafin. 1991.Cloning of the cDNA and gene of mouse mast cell protease-6. J. Biol. Chem.266:3847-3853.

40. Serafin, W. E., D. S. Reynolds, S. Rogel, W. S. Lane, G. A. Conder, S. S.Johnson, K. F. Austen, and R. L. Stevens. 1990. Identification and molecularcloning of a novel mouse mucosal mast cell serine protease. J. Biol. Chem.265(l):423-429.

Peptidase Modulation of Vasoactive Intestinal Peptide Pulmonary Relaxation 243